1
$\begingroup$

This question already has an answer here:

So I'm having some troubles with a signal that's been sampled but where the samples don't have consistent intervals. I've been looking into the NFFT (in python) as a starting point but I'm a complete noob with signal processing so I'm not really getting anything out of it.

So I have one array with the sample values and another with the corresponding timestamps. But I really have no idea how to get the spectrum out of it. I've done the first part as described here.

Sorry about the sparse question.

$\endgroup$

marked as duplicate by MBaz, Tendero, Matt L., lennon310, A_A Sep 5 '18 at 14:54

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

  • $\begingroup$ Hi! So if you have the nonuniform samples and the associated sampling time instants, then all you have to do is to feed them into the so called nonuniform FFT function. Please consult into Python help system, for learning how to do it in the software. $\endgroup$ – Fat32 Sep 4 '18 at 15:07
  • $\begingroup$ There are quite a few related questions on this website: take a look at dsp.stackexchange.com/q/22253/11256 dsp.stackexchange.com/q/25524/11256 for example. $\endgroup$ – MBaz Sep 4 '18 at 15:19
  • $\begingroup$ Tnx for the comments, I'll take a look! $\endgroup$ – RFmyD Sep 4 '18 at 15:36
1
$\begingroup$

The problem you have to solve is having a model for how your samples uniquely represent actually continuous input. For equidistant samples, the usual model is to stipulate (and then guarantee by analog filtering, oversampling, digital filtering, and decimation) that there are no signal components exceeding the Nyquist frequency.

What is your model going to be for the kind of non-uniform sampling you are working with? If the information you work with is not sufficient for that model, you are going to have non-uniquely interpretable results, namely aliasing.

$\endgroup$

Not the answer you're looking for? Browse other questions tagged or ask your own question.